The Production and Prediction of Traffic Accidents

Traffic accidents are a painful and costly affliction all over the world. The discussion and debate over the most effective measures to adopt to enhance road safety is often based on folk theories and is not free of private interests and pressures. Road traffic is a complicated system of interactions providing transportation services. Unfortunately, in conjunction with these services, traffic accidents, an awful public bad, are produced. The purpose of this paper is to study the production relationship, for intercity transportation, between the physical factors, i. e., highways and vehicles, and traffic accidents. The upshot is prediction ability of accidents in any given road segment, existing or planned. The regretful aspect of roads runs counter to conventional wisdom which, failing to appreciate the quantitative relationship between roads and accidents, often advocates building more and wider roads as the remedy. The empirically substantiated public bad property of roads, by way of production functions for traffic accidents, is useful for public policy, concerning the investment in highways versus other forms of transportation, such as rail. They also promote better understanding of traffic accidents and their data and the methodology allows testing hypotheses relating to safety policy. This study sheds light on the enigma of the long-term decline in the probability of death on the road, as observed in many countries, by attributing it to the rising traffic density. The estimation also sheds light on the accident-externality imposed by road users on others. The results suggest that for inter-city roads the risk of fatal and severe accidents is over-internalized by road users as the marginal effect of traffic flow is smaller than the average.